C++ Library for Variational Monte Carlo simulations. It is built upon the MCI++ (https://github.com/DCM-UPB/MCIntegratorPlusPlus) and NoisyFunMin (https://github.com/DCM-UPB/NoisyFunMin) libraries. It also uses the GSL (GNU Scientific Library).
In doc/ there is a user manual in pdf (not accurate for current master!) and a config for doxygen.
In examples/ and test/ there are examples and tests for the library.
Some subdirectories come with an own README.md file which provides further information.
Currently, we automatically test the library on Arch Linux (GCC 8) and MacOS (with clang as well as brewed GCC 8). However, in principle any system with C++11 supporting compiler should work.
- CMake, to use our build process
- MCI++ and NoisyFunMin (master)
- GNU Scientific Library (~2.3+)
- (optional) a MPI implementation, to use parallelized integration
- (optional) valgrind, to run
./run.shintest/ - (optional) pdflatex, to compile the tex file in
doc/ - (optional) doxygen, to generate doxygen documentation in
doc/doxygen
Copy the file config_template.sh to config.sh, edit it to your liking and then simply execute the command
./build.sh
Note that we build out-of-tree, so the compiled library and executable files can be found in the directories under ./build/.
You may want to read doc/user_manual.pdf to get a quick overview of the libraries functionality. However, it is not guaranteed to be perfectly up-to-date and accurate. Therefore, the best way to get your own code started is by studying the examples in examples/. See examples/README.md for further guidance.
This library supports multi-threaded MC integration with a distributed-memory paradigm, thanks to Message Passing interface (MPI).
To activate this feature, set USE_MPI=1 inside your config.sh, before building. The header MPIVMC.hpp provides convenient functions
for using VMC++ with MPI. For example usage, look into example ex7.